Nowadays the use of aerobic granular sludge, also known as Nereda™ in municipal wastewater treatment, is applied worldwide. The next innovation in for this technology is the implementation of granular sludge in continuous flow through configuration. However, many challenges will have to be faced before this technology can be applied in full scale wastewater treatment plants. One of these is the formation and behaviour of granular sludge in continuous flow reactors. In this study the start-up and steady state behaviour of two lab-scale bioreactors were compared on growth speed, granule size distribution, influent distribution and density. One of the reactors was run with an up flow feeding through the sludge bed, as performed in Nereda™ plants. The second reactor was run with a pulse feeding, followed by an anaerobic mixing phase. By combining data from image analysis with reactor cycle measurements, batch tests, PHA determination and density measurements, a classification of the granule size distribution and COD distribution over the granules was made. A comparison between the reactors could be made, as both reactors were exposed to the same conditions, except for their anaerobic influent regime. Both reactors were fully granulated at the end of reactor operations. The bed fed reactor had a faster average growth speed of granules due to the higher COD load on the biggest granule fraction. This selected for a higher mass fraction of granules greater than 1mm in diameter in the bed fed reactor. Both reactors seemed to suffer from diffusion limitation, as oxygen limitation was observed in the bed
fed reactor, while the pulse fed reactor showed spatial COD limitation. Both reactors experienced some degree of P-limitation. The results from this study can be used to model granulation and granule metabolism of aerobic granular sludge exposed to different anaerobic contact regimes.

Treatment of wastewater with aerobic granular sludge (AGS) has major advantages compared to activated sludge. The only commercial system that uses AGS is the Nereda® batch process. The disadvantage of this system is the lack of existing infrastructure. AGS in continuous systems could be a solution to this disadvantage. That’s why the HARKOS pilot is built. This pilot is based on the enhanced biological phosphorus removal (EBPR) process. The aim of this study is to make sure this pilot runs according to the EBPR process (reactor characterization) and to apply biological selective pressure for granulation. This biological selective pressure consists out of four pillars. Prevention of COD overshoot into the aerobic zone, as much COD in the anaerobic zone to storage polymers, enrichment for bio-p organisms and compartmentalization of the reactor. It’s concluded that the pilot is characterized, that no COD overshoot is present, that more COD is stored as storage polymers, that the sludge is enriched for bio-p organisms and that the reactor is well enough compartmentalized. The biological selective pressure is applied. If enough physical selective pressure is applied granulation will start in the HARKOS pilot.